MIT EECS Undergraduate Research and Innovation Scholar
Low-power Processor for Realtime Motion Magnification in Videos
Anantha P. Chandrakasan
Videos can provide information about small movements that the human eye would not normally perceive. Motion magnification techniques have made visualizing these small, periodic motions possible. Previous motion magnification algorithms focus on processing the entire video frame to produce a motion magnified video. This project focuses on incorporating motion detection into these techniques to reduce the size of the processed video, thereby reducing runtime and energy required for processing.
My previous research experience focused on motion planning and computer vision on PR2 robots. I have been working in MTL under for the summer and have familiarized myself with the previous work and explored future applications of the technique. I am excited to contribute to the improvement of the technique.